Combining Behavioral Endocrinology and Experimental Economics: Testosterone and Social Decision Making

This video demonstrates an experimental economic paradigm that can be used to show the impact of exogenously administered testosterone and the role of expectations about the hormone in human bargaining behavior. First, female students are recruited from an online database. Testosterone or placebo is randomly administered to the subjects in a double-blind manner, artificially raising testosterone levels in the group, receiving the verum, and leaving testosterone levels at baseline in the placebo group.

Subjects then play three rounds of an economic bargaining game in order to assess fairness of decisions during actual social interaction. Finally, statistical tools are used to analyze effects of testosterone administration together with the effects of expectations about the hormone. The following experiment requires a considerable amount of control over the study environment.

Up to 12 subjects may be present in the laboratory at the same time, and communication between subjects is strictly forbidden. Experiment or demand effects should be avoided by minimizing direct personal contact with the participants. Finally, anonymity should always be explicitly assured to reduce effects of social desirability, human subjects for experiments.

Combining behavioral endocrinology and experimental economics are recruited through an online database containing a list of a large number of volunteers from this list. Only women who have not previously participated in similar studies are contacted. It is important to recruit women only because the time course of neurophysiological effects after acute testosterone administration is only well established among women.

For men this time course is not well understood. Interested subjects are then screened in a telephone interview to exclude those using hormonal contraception, suffering from somatic diseases or any neurological or psychiatric disorders. Because endogenous levels of sex hormones tend to be low and stable within the first part of the female menstrual cycle.

A second telephone interview is conducted to gather information about regularity and duration of the subject cycle. Subjects are then invited to come to the lab for the experiment in groups of 12. Within 10 days of the beginning of their menstrual cycle, they're further instructed to abstain from alcohol or caffeine intake and smoking 24 hours.

Preceding the experiment. For this experiment, a classical behavioral decision making laboratory is needed, which has sufficient individual computer workstations to run experimental sessions of a group of subjects. In parallel, workstation should be connected through a network and run the appropriate presentation software.

Importantly, the laboratory needs to be arranged so that visual and oral communication between subjects is impossible or very limited, and so that the experimenter can control and observe all subjects. Due to the interdisciplinary nature of this experiment, the behavioral laboratory must have basic biomedical equipment nearby, such as a standardized specimen freezer. In addition to the standard laboratory infrastructure, each workplace is equipped with an appropriate kit for saliva specimen collection.

Additionally, a glass of water is placed at each workstation. In preparation for the experiment, a person who will not have personal contact with subjects randomly distributes the testosterone and the visually indistinguishable placebo pills to the numbered workstations to control for the well-known circadian variation in steroid hormone levels begin all sessions at the same time of day. Here sessions begin at 1:00 PM when the subjects first enter the laboratory, have each of them draw a card with a concealed number on it, which corresponds to the numbered workstations.

Then take the subjects to the laboratory room and instruct them to seek out and sit down by the workplace with the same number as indicated on the card they drew at the entrance. This allows for a random allocation to the drug and placebo treatment. Once subjects are seated, let them read and sign the consent form and the form by which they confirm that they are not pregnant.

Collect all the signed forms and instruct the subjects to use the saliva collection kits and then immediately freeze the samples at minus 80 degrees Celsius. This will be used to measure their baseline salivary testosterone and cortisol concentrations. Next, have the subjects place the sublingual pill into their mouth, chew it and keep the disintegrated substance in their mouth for one minute.

To achieve maximal contact of the hydroxy propyl Beta Cyclodextrin Testosterone inclusion complex with the oral mucosa, subjects should make circling movements with their tongue after one minute, instruct the subjects to swallow the remainder of the substance. With the 100 milliliters of water provided, immediately following administration of the substance have the subjects complete questionnaires on screen to control for any effects of subjects, mood and personality. In this experiment, mood and arousal are measured with the multidimensional mood questionnaire and anxiety and anger.

States are measured with the state trait anxiety inventory and the state trait anger expression inventory due to the established lag time of four hours for the appearance of behavioral effects. After sublingual application of testosterone and female patients, a long waiting period is an integral part of the study. During this period, subjects are required to stay in the laboratory room and keep themselves quietly occupied.

This ensures that no social interaction outside the laboratory takes place, which could give rise to confounding measures. When the waiting period has passed, instruct the subjects again to use the saliva collection kits and then immediately freeze the samples at minus 80 degrees Celsius. Have the subjects complete questionnaires on screen.

Again, following this, the ultimatum game can begin four hours following pill administration randomly and anonymously. Assign each subject to the role of the proposer or the responder. Provide paper instructions corresponding to their role in the game and the payoff rules.

Check their comprehension of the payoff structure by having them complete several questions after reading the instructions. Next orally, summarize the experimental procedure to all participants at once. This ensures that subjects know that everyone has the same information about the structure of the game and how the payoffs are determined.

All proposals and decisions in the ultimatum game are implemented in Z Tree software and presented on computer screens. The ultimatum game starts with the proposer's decision. She has to decide on how many of the 10 monetary units or MUS to offer to the responder.

She can choose to offer either 5, 3, 2, or zero MUS to the responder. In this example, she chooses to offer the equal distribution IE to offer five mus by ticking the corresponding box on the screen. This decision is then transmitted to a randomly selected responder.

The responder now sees the proposition and is asked to decide whether to accept or reject this proposal. In this example, the responder accepts the proposal. Thus, in this example, both the proposer and the responder would earn five mus.

This game is played three times each time the group composition has changed. However, the decision on the responder is only communicated to the proposer after the proposers have made all three propositions. This ensures that all the proposers are in a similar situation for all three choices.

Since feedback on the responder's reaction might influence proposers subsequent choices otherwise, now a summary screen with their earnings in the experiment is displayed. In this example, the proposer twice offered five and once two mus to the three different responders. The former offers were accepted and the latter rejected resulting in a total payoff of 10 mu.

For the proposer. Because testosterone is widely discussed in the popular press, subjects may have a preformed belief about its effects. Furthermore, because ethical concerns require researchers to inform subjects that they will either receive a placebo or testosterone, it is important to determine if the subjects think they have received testosterone or the placebo pill by explicitly asking them on screen.

This perceived group assignment is recorded in the final questionnaire together with a few socioeconomic characteristics. When all the final questionnaires are completed, the subjects are asked to prepare for leaving and wait until called upon for payment. The payment takes place individually and in a separate room to ensure privacy and anonymity.

The payment consists of a flat fee of 100 Swiss francs plus a varying amount depending on the outcome. In the ultimatum game. As an all economic experiments, it is important to set financial incentives.

Otherwise, people might behave fairly simply because it looks good to others and does not cost them anything rather than because they really care about the others'benefits. By making the payoff contingent on the decisions in the ultimatum game, we can make fair behavior costly for the proposer, and thus ensure that the behavior reflects what she really cares for. After payment subjects are free to go.

Results show that the experimental procedure was successful as subjects who receive placebo and those who believed they had received the placebo made decisions that are in line with standard ultimatum game behavior. The green bars on the left show the distribution of offers that are typically observed in the ultimatum game. The horizontal axis represents the proposer's offers in the ultimatum game, and the vertical axis depicts the frequency with which each of these offers is observed.

The blue bars show the distribution of the results qualitatively. Both distributions look similar, providing evidence that our procedure is successful. Furthermore, the approach described here has proven to be sensitive to pharmacological challenges as subtle as those achieved by testosterone and administration.

On the left graph, you can see the effect of testosterone on proposer's average offer. Subjects who receive testosterone make higher offers than subjects who received a placebo, an effect which is significant when controlled for expectations about the hormone. This additional important observation is shown in the right graph.

The offers from subjects who believe to be in the placebo group are higher on average than the offers from the subjects who believe to be in the testosterone group. Thus, it seems that the beliefs about the administered substance itself can have a powerful influence during social decision making situations. The main advantage of this procedure of our existing methods is that we artificially raise subjects as the shown levels and are therefore able to make causal inferences about the hormones effect on behavior.

A further advantage is that we directly observe decision making behavior rather than any form of self-reported behavior. It is important to acknowledge the truly interdisciplinary nature of this project for successful implementation of this experimental procedure, this essential to ensure a tight cooperation between experts in the corresponding fields.